Today Fireship covered this: - Indie hackers may be the first type of developer to go extinct - The classic indie hacking playbook (last decade): learn to code → find a niche → ship a micro-SaaS → build in public → post MRR screenshots after failures → repeat until... it works - Coding used to be hard and expensive, so execution mattered far more than ideas - That equation has flipped: execution now costs ~$20/month, so anyone can build personal software instead of paying for SaaS - Example: Why pay $29/month for a product when Claude Code can recreate a better version in 20 minutes? - Building in public on X makes it easy for others to steal your idea/roadmap - AI raises the ceiling (indie hackers become far more productive) but also lowers the floor (the space is now crowded with normies who can vibe-code their own tools) - This trend will accelerate as AI gets cheaper and smarter - Silver lining: the disruption can "fertilize the soil" for those who master distribution, branding, or proprietary datashow more

@fireship_dev why not just build a great long term project instead of crappy saas to make some short term revenue ?

@fireship_dev I don't think most indie projects are crappy at all I do think they need more/different distribution and more complexity or things that AI can't vibe code quickly

@Xmorosini @fireship_dev How is this metrially different from before? If you spend the same amount of time and effort on something you can now ship much bigger projects? It's not low-hanging fruit, but... Why is the playbook broken if you adjust your aim upward?

@Xmorosini @fireship_dev Because before it took 5 weeks/months to make something you can now make in 5 minutes Bigger projects isn't necessarily better right? It's unclear what is better though for me, it's a bit of an odd transitionary time

@Xmorosini @fireship_dev I don't think that's accurate. I must be doing it very wrong if everything that took five months is achievable in five minutes. Because it just is not.

@Xmorosini @fireship_dev Have you tried Opus or Fable on Ultracode yet?

@fireship_dev I think indie hackers should probably start building products for more sophisticated niches where you still need to have technical skills and vibe coders can’t reach yet. To-do lists and similar app are no longer relevant.

@fireship_dev Yes and niches that tie into IRL businesses Not just only online

indie hackers weren’t getting paid to code, they were getting paid to be early & to build opinionated products that corporations designed by committee can’t build I don’t think the entire purpose dies, but it’s gonna look WAY different. @mattshumer_ ‘a recent AAA call of duty loop becoming a meme feels a lot closer to the future than launching a SaaS. The trick here is the entire business model is shifting or maybe even gone. The new model is front running companies and making software memes.

@fireship_dev @mattshumer_ Yes I agree with this Parallel with musicians making 90% of their money with songs before internet Now they make 90% with shows

@fireship_dev @mattshumer_ Yessss great analogy. Is that why you started your merch store?

@fireship_dev @mattshumer_ Yes trying anything, mostly tweeting ChatGPT in 2022 told me to go hard on building a big audience cause it was the one thing AI couldn't replace so I did

@fireship_dev @mattshumer_ Oh dam. I didn’t realize you were focusing on that. That chart is insane. Kudos. Well done.

@fireship_dev You are saying people rather work and spend countless hours chatting with it just to create a tool? People want convenience. There will always be demand for a ready made tool. When the problems get easier, people also get lazier.

@fireship_dev Now yes but very soon it'll take 200 miliseconds to spawn an app that does same as majority of SaaS

@fireship_dev There is still the human brain barrier man, getting information out is a hassle.

@fireship_dev Congrats on making it to fireship! Best channel ever :p

@fireship_dev I agree 😊 I was just downloading its videos for a flight in EU with no WiFi

@fireship_dev 😂

@fireship_dev maybe not go extinct, but software alone is no longer defensible, the only 2 moats left are: - information asymmetry - trust I restructured my company recently because of this:

@fireship_dev Yeah, 2y ago no dev would exist today. Now no “indie hacker” will exist. Indie hacker is a symptom of capitalism, not technology. While still has business needs weird needs, people will pay for it. Maybe selling for devs is harder now but that’s not true for most industries

@fireship_dev My only question, if AI is getting so much better and smarter, why wouldn’t it also take a lot of the distribution up as well? What are your thoughts on that?

No. No. No. Coding was indeed expensive but it was just a component of software products, being them apps, SaaS, etc. There is a lot more backed in: customer support, maintenance, marketing. Solving one process will create some asymmetrical opportunities for a specific segment, but it won’t eliminate an entire “population”. The asymmetrically advantaged cohort is quite thin: good idea + execution + maintenance. Oh, and the “I just vibecoded this paid product” is just “I wrote an entire website without writing a line of HTML”. The doomerism is getting annoying.

@fireship_dev I think you may underestimate how lazy the common man is and how much he is willing to pay just so he doesn't have to prompt because he still doesn't understand how any of this works, and probably never will. This is true from small businesses to enterprise customers.

@fireship_dev Indie hackers should build on @lukso_io , it's a new edge that puts digital identity at the center of everything.

@fireship_dev Pieter levels leading upcoming generation of indie hackers astray. One only needs to look at the' starter story' YouTube channel to see that indie devs are increasing and not decreasing!

@fireship_dev Don't shoot the messenger

@fireship_dev Apologies for the harshness in the first post, but I absolutely do want to shoot down the message. No disrespect to you, or belittling of your achievements. As a suggestion you could caveat these arguments more often, to offset the fear-inducing factor of such a message

@fireship_dev Agree the floor dropped. But $20/month covers building the thing. It doesn't cover the support inbox, the chargebacks, or being the one who fixes it when it breaks. That side never got cheaper.

@fireship_dev the "$29/mo vs 20 minutes with claude code" math always skips maintenance. teams don't pay for features. they pay so it's not their problem at 2am.

@fireship_dev the cost everyone forgets is time. aint nobody gonna waste their time to vibecode everything. they want to spend their time on their own thing or having fun.

@fireship_dev You've inspired more devs to become indie hackers than anybody else, so this is more than just an observation -- more like a self-fulfilling prophecy.

@fireship_dev Yes it's naturally things change and my point is people have to surf the waves

@fireship_dev 100% i am just adding that some of the waves people were riding were probably just your wake

@fireship_dev Not just distribution but becoming a product scientist And optimizing onboarding, and making data driven decisions There’s still a need for infrastructure products and products that work with Claude

@fireship_dev but still we are 8b on this planet. you can always find people to buy your stuff. For sure it will be much harder and you might not be able to make so much money.

@fireship_dev I'm almost happy that I've got <$1000/MRR with indie hacking 😂 I've got nothing to lose

@fireship_dev Silver lining: the disruption can "fertilize the soil" for those who master distribution, branding, or proprietary data Pretty much yeah

@fireship_dev In an ideal world anyone with a great idea should have a means to execute it, so in my opinion the playing field becoming even (re execution) will benefit those who deserve it most.

@fireship_dev Would you say that building in public is dead? Like why are you gonna share info that is gonna be useful for your competitors?

- A dev will vibe code a replacement for a dev tool. A consumer will not vibe code their own iOS app and release it to the app store - $20 Claude code credits is not enough to create fully featured app - AI still very much creates inferior UX to someone with good UX instincts - Many non-techies don’t know that ChatGPT or Claude can create a whole software for them - People are lazy and want something that works. Just as a large percentage of Gen Z eats out instead of cooking at home, vast majority of people prefer to pay a premium for good software that works as opposed to going through the headache to create their own

@fireship_dev - Example: Why pay $29/month for a product when Claude Code can recreate a better version in 20 minutes? Because your own time prompting Claude is worth more than $29/month. Infra, upkeep not even taken into account.

@fireship_dev I think there is still opportunity for a micro saas in a niche but that niche needs to have some gating element to it. Eg think building software for your local hospital where you need to haggle them to give you access to proprietary hospital license just to dev

@fireship_dev You can create music with Suno but most people want music from an artist. Most of suno's subscribers make music to publish on Spotify. Most people just don't want to take the time to build something. Especially when the cost is higher to host than to just buy.

What GPT-5.6 SOL thinks about the youtube video: ====== AI does not make indie hackers obsolete. It makes “I can code this” far less valuable as a differentiator. The vulnerable category is the simple, low-risk, single-user SaaS whose main value was saving someone from writing code. The stronger category sells maintained outcomes: trusted data, recurring workflows, collaboration, integrations, reliability, privacy, auditability, support, and responsibility. The deepest principle is: AI makes production abundant. Therefore evidence, trust, access, maintained context and responsibility become more valuable. What the video gets right—and wrong:- Anthropic’s own research supports the core concern: 79% of sampled Claude Code conversations involved automation, and simple user-facing applications appear especially exposed. Anthropic’s software-development analysis But “execution now costs $20/month” is misleading. That buys access to a capable coding tool. It does not buy: - Correct product judgment - Reliable data pipelines - Hosting, maintenance, monitoring and security - Integrations and permissions - Regulatory accuracy and version control -Distribution and trust operational responsibility Distribution and trust Even Anthropic positions Opus 5 around stronger coding, verification and long-running work—not autonomous ownership of a commercial product. Claude Opus 5 announcement The market is likely to split: Increasingly vulnerable > Increasingly valuable: - Generic wrappers > Maintained workflow systems - One-off personal utilities > Recurring team workflows - Commodity public information > Normalized, versioned, source-backed data - “Chat with this PDF” > Reviewable work across many documents - Low-risk convenience > High-stakes accuracy, privacy and auditability - Features that can be regenerated > Embedded relationships, integrations and trust The middle—ordinary micro-SaaS with shallow differentiation—will get squeezed. But software businesses do not disappear merely because software becomes easier to produce. Most customers do not want to become software operators. Distribution will also be automated—but durable distribution remains: One X commenter is right that AI will automate parts of distribution: - Content creation - Outreach preparation - SEO research - Ad production - Personalization - Lead qualification That means generic marketing output becomes noisier and less valuable. What remains scarce is accumulated distribution: - A permissioned audience - Trusted personal or product brand - Customer references - Direct industry relationships - Integrations and marketplace placement - Original data and authoritative research - Being the default source an AI agent consults AI also creates a new distribution surface: AI search and software agents choosing sources and tools for users. Google’s guidance still emphasizes useful, reliable, original and non-commodity content rather than a separate “AI SEO” trick. Google’s AI-search guidance Anthropic’s Model Context Protocol similarly illustrates how products can become tools and data sources used by agents. Model Context Protocol So distribution does not disappear. Some of it moves from “convince a human to click” toward “be the trusted, structured source a human or agent selects.” Building in public after AI: The video overstates the roadmap-theft risk. Ideas and visible features were rarely strong moats. Still, your public-sharing policy should evolve: Share: - Shipped outcomes - Real examples - Product principles - Customer problems, when anonymized Specific prospects and commercial negotiations: Build in public should create trust and distribution—not provide a live specification for unfinished work. The deepest principle is: AI makes production abundant. Therefore evidence, trust, access, maintained context and responsibility become more valuable.

@fireship_dev building it yourself is free until it breaks on a saturday and you find out you are also the support team

@fireship_dev But that Idea which is getting executed by building in public is suggested by AI. The innovative idea still stands different and is eligible to generate millions of dollars in revenue. But the problem is that the idea is not reaching its user due to high competition.

@fireship_dev As a normie who vibe codes his own apps, I have to agree. Build a proper moat or get raided by the savage hordes.

@fireship_dev distribution was always the boss fight, AI just skipped the tutorial :D

@fireship_dev If you are building hardware, you can be safe as an indie hacker. It is really hard to make vibe coded hardware still. It does work, but "normies" will nib be able to do it without putting effort, unlike with vibe coding.

@fireship_dev Feels like 1994 again, when internet started reshaping all types of businesses

@fireship_dev The tell in levelsio's thread is that traffic dropped before revenue. Micro-SaaS ran on long tail search queries, and those get answered now without anyone clicking through. Everyone is arguing about AI building competitors when the distribution layer went first.

@fireship_dev want to highlight the most important part of this post: DISTRIBUTION. BRANDING. PROPRIETARY DATA. this is what you need to focus on if you want to survive.